Here is what I'm getting (needed to test this anyway, thanks for asking).
EXT3 about 1.5Mbytes/sec:
adsdebian:/mnt/hda# df . Filesystem 1k-blocks Used Available Use% Mounted on /dev/hda2 1027800 16428 959164 2% /mnt/hda adsdebian:/mnt/hda# time dd if=/dev/zero of=foobar.ext3 bs=128k count=1024 1024+0 records in 1024+0 records out
real 1m21.735s user 0m0.000s sys 0m33.680s
VFAT about 1.5Mbytes/sec:
adsdebian:/mnt/hda# df . Filesystem 1k-blocks Used Available Use% Mounted on /dev/hda2 1042172 4 1042168 0% /mnt/hda adsdebian:/mnt/hda# time dd if=/dev/zero of=foobar.vfat bs=128k count=1024 1024+0 records in 1024+0 records out
real 1m24.010s user 0m0.040s sys 0m13.470s adsdebian:/mnt/hda#
Does anyone have any empirical evidence as to the ep93xx IDE throughput available using the latest linux-cirrus-2.6.x kernel? Something along the lines of 'time dd if=bigfile of=/dev/hda'?
I realize that DMA is not (yet) functional; I assume PIO modes 0-4 all work though?
Just trying to get a handle on an upcoming project. Thanks for any and all input.
-Cory
-- Robert Whaley Applied Data Systems www.applieddata.net 434-244-9504 rwhaley@xxxxxxxxxxxxxxx